Baas-Becking for Trouble?

I am currently at my first conference (The GRC on Applied & Environmental Microbiology), and I thought that writing about some of the topics being covered at it would be good. Last night, the conference opened with a discussion of the ubiquity of microbe "species", so I thought that would be a good place to start here too.

Perhaps one of the most provocative statements made by a microbiologist to date is Baas Becking's 1934 statement that "alles is overal: maar het milieu selecteert". This translates to "Everything is everywhere but the environment selects", which I interpret as through wind and wave, microbes have the ability to disperse anywhere on the planet, although whether or not they are able to thrive and grow depends on their needs.

 You may think this sounds a bit obvious - how could something like Neisseria gonorrhoeae, the human-dependent bacteria which causes gonorrhea, be found surviving and thriving in Antarctica, thousands of miles from the nearest human? And we know that the obligate pathogen Bacillus anthracis, which causes anthrax, is not in most people's lungs, because if it was, they would be dead.
And yet this question is under particularly intense debate at the moment. But why?

To understand this question, we have to consider the journey environmental microbiology has taken since Becking made this statement. Twenty or thirty years ago, if you wanted to know whether your microbe of interest was everywhere, you would have to take samples of water or soil or rocks from all different places, and then use a series of different growth conditions to try and enrich for the microbe. But not all microbes are culturable using current techniques, or perhaps at all, so we are missing out on some of the picture. But perhaps more importantly, a given microbe may be very easily cultivated and identified in samples from some places, and present but impossible to culture from another place, meaning that it may be present - even thriving - but not detected. Shockingly, this is the case for some strains of fecal bacteria used as indicators of water quality - they may become unculturable after passing through wastewater treatment, making it difficult to assess the safety of the effluent. 

Fortunately, the rapid growth of new culture-independent methods for detecting microbes in the environment has ensured that our understanding of who is present (though not always what they are doing) is much less of a problem. In these methods, researchers collect a piece of the environment (seawater, soil, leaves, rocks), extract the DNA, and using microbe-specific DNA fragments as primers to intiate sequencing reactions, they sequence the DNA (usually just the ribosomal RNA sequences) in the sample. These rRNA sequences are like barcodes for the microbes, and the most widely-used definition of bacterial "species" is based on similarity of this sequence. Thus by sequencing just a short stretch of DNA, we can see who is present, and theoretically we can detect any microbe that is present whether or not we know anything about its preferred growth conditions. This makes it much easier to answer the "everything is everywhere" question! Theoretically.

Even with this enhanced ability to see who is where, we are still debating whether everything is everywhere, possibly because, as my PI pointed out, this question means something very different to the day it was first made. In a study examining seasonal changes in the microbial community at a site in the English Channel, it was noted that organisms previously thought to be lost from the community were in fact present in low numbers and possibly deeper in the water column; this is the so called microbial seedbank hypothesis. Microbes are everywhere, in low abundance in various stages of dormancy, and the environment selects from this pool. Here, time (everywhen) is used as an analog for everywhere.
But other studies have concluded that in fact, everything is not everywhere. For example, in a study utilizing data from thousands of samples taken all over the global oceans, researchers from Woods Hole found that in some instances, geographical proximity, rather than environment type,  dominates whether a given microbial species is present. The environment selects, but dispersal limitations also play a key role in this.

Some researchers have responded to the observation that everything is not everywhere by stating that not all microbes are found in all samples because "sequencing isn't deep enough" - that is, not enough DNA has been sampled and sequenced from the environment - and some of the "singletons" (or sequences which appear only once in a sample) which are routinely assumed to be sequencing errors and therefore discarded may be real, though rare organisms. Furthermore, there is always at least some sequencing bias - the primers used to initiate sequencing runs may not bind all bacterial genomes equally or at all, meaning that some microbes are missed. Even if we could sequence all kinds of bacteria, the depth neccessary would come at a high cost: Tim Vogel of the Ecole Centrale de Lyon estimated that we would need about a thousand Illumina sequencing runs to get all the microbes in a gram of soil, which would cost in the millions of dollars per sample. Of course, as many of the speakers and commentors brought up last night, a much cheaper way to prove that everything is everywhere is to define everything at a broad taxonomic level (ie bacteria vs. E. coli 0157:H7 substrain xxx) and everywhere at a large scale (for example, on this continent)!

So after all that, is it worth it to try and sequence all these bacteria to see if everything really is everywhere? Last night, Rob Knight told us that knowing that everything is everywhere is very important for understanding how to treat patients. Take a patient about to undergo chemotherapy that will wipe out his immune system. If a potential pathogen is already in his body, putting him in a clean room will do little good. But if it isn't, then taking this preventative approach could save his life. But in other systems such as soils, there appears to be sufficient microbes doing the same thing, and stochastic processes driving the extinction and local recolonization of species, that everything being everywhere isn't a particularly good (or "biologically informative") debate to have. Maybe we need to look at whether functions - rather than arbitrarily defined species - are everywhere, and what functions the environment selects for.